Video stitching is a technology to stitch several videos which have overlapped parts into a large-scale, high resolution, seamless video.
Traditional ways to realize video stitching mainly include:
1) A fixed number of cameras are fixed in a closed apparatus, the number of the cameras is at least two, and angles between the respective cameras are fixed; since in this way, the number and angles of the cameras are all fixed, for different scenes, the adaptability and flexibility is relatively poor.
2) Separate cameras are used and mounted respectively, and similarly, the number of the cameras is at least two, and the number and angles of the cameras may be selected according to requirements of real scenes; though in this way, the adaptability and flexibility is relatively good, the overall procedure of mounting and testing the cameras all need to be done manually, and after the procedure of mounting and testing is completed, if image positions of the cameras move due to vibration, wind and rain, etc., the cameras need to be tested again, which is troublesome to implement, and costs a great amount of manpower and time.